JP3785195B2 - Method for constructing the molecular structure of a biologically active ligand - Google Patents

Description

次に、Ｃ５において、上記のＣ４のステップで抽出した官能基を構成する原子の中からコネクターを結合させる２個の原子の組み合わせのすべてを設定する。例えば、＞ＮＨと−ＯＨをＣ４のステップで選択した場合には、＞ＮＨのＮ原子および−ＯＨのＯ原子を選択する。ここで、コネクターとは、Ｃ４のステップで抽出した官能基をつなぐ一連の原子集団をいうものとする。
【００４５】
次に、Ｃ６において、Ｃ５で設定した２個の原子の組み合わせの中から１つを選択する。ここで選択された２個の原子をATOMhb1 および ATOMhb3と称することとする。また、このATOMhb1 に付加するダミー原子をATOMhb2 、ATOMhb3 に付加するダミー原子をATOMhb4 と称することとする。ここで、ダミー原子とは、コネクターの適否の判定のために、付加する仮想的な原子をいうものとする。ダミー原子ATOMhb2 およびATOMhb4 の位置は、抽出した官能基の位置と方向によって必然的に決まってくる。以下、ATOMhb1 〜ATOMhb4 を官能基側の原子ATOMhb1 〜ATOMhb4 と称することとする。
【００４６】
次いで、Ｃ７において、Ｄ値の限界値、およびＴ値の限界値を設定する。
Ｄ値は下記の式（１）から算出される値で、官能基側の４つの原子ATOMhb1 、ATOMhb2 、ATOMhb3 およびATOMhb4 の距離関係とコネクター側の４つの原子ATOMcon1、ATOMcon2、ATOMcon3およびATOMcon4の距離関係の類似性を示すものである。ここで、ATOMcon2およびATOMcon4はコネクターの両端の原子を表し、ATOMcon1およびATOMcon3はコネクターの両端に付加されるダミー原子を表す。
【００４７】
Ｄ値＝（Ｄ１）² ＋（Ｄ２）² ＋（Ｄ３）² ＋（Ｄ４）² 式（１）
（式中、Ｄ１は（ATOMhb1 とATOMhb3 の距離）と（ATOMcon1とATOMcon3の距離の差を、Ｄ２は（ATOMhb2 とATOMhb4 の距離）と（ATOMcon2とATOMcon4の距離）の差を、Ｄ３は（ATOMhb1 とATOMhb4 の距離）と（ATOMcon1とATOMcon4の距離）の差を、Ｄ４は（ATOMhb2 とATOMhb3 の距離）と（ATOMcon2とATOMcon3の距離）の差を表す。）
Ｄ値が小さい程、それぞれの４原子の距離関係が類似していることになる。
【００４８】
Ｔ値は下記の式（２）から算出される値で、官能基側の４つの原子ATOMhb2 、ATOMhb1 、ATOMhb3 およびATOMhb4 のねじれ角とコネクター側の４つの原子ATOMcon2、ATOMcon1、ATOMcon3およびATOMcon4のねじれ角の差を示すものである。
Ｔ値＝Ｔ_hb−Ｔ_con 式（２）
（式中、Ｔ_hbはATOMhb2 、ATOMhb1 、ATOMhb3 およびATOMhb4 のねじれ角を、Ｔ_con はATOMcon2、ATOMcon1、ATOMcon3およびATOMcon4のねじれ角を表す。）
Ｔ値が小さい程、それぞれの４原子の位置関係が類似していることになる。
【００４９】
Ｄ値の限界値およびＴ値の限界値は、ユーザーが自由に選択することができるが、Ｄ値が２〜５オングストロームの範囲で、Ｔ値が１０〜２０°の範囲であると、精度よくかつ効率的に適切なコネクターを選択できるので好ましい。
次に、Ｃ８において、フラグメント集団を用意する。フラグメントとは、コネクターを作成するための部分構造をいうものとする。フラグメントとしては、
４価の炭素（ＳＰ³ のＣ）等の原子、およびベンゼン環、エチレン、カルボニル、シクロヘキサン等の官能基を使用してよい。
【００５０】
次いで、Ｃ９において、コネクター内の回転可能な結合の回転様式を設定する。大抵の場合、コネクター内には回転可能な結合が存在するので、それらを系統的に回転させることにより、官能基間にうまく嵌め込めるコネクターのコンホメーションを見出すことができる。例えば、回転可能な単結合の場合には、２０°〜６０°のステップでねじれ角を系統的に変化させる回転様式を設定し、回転可能な結合が環構造内にある場合には、可能な種々の環構造を順次ファイルから入力できるように設定することができる。
【００５１】
次に、Ｃ１０において、Ｃ８で用意したフラグメント集団のフラグメントをすべて組み合わせて種々の長さのコネクターを作成し、これらを保存したテーブルを用意する。テーブルとしては、短いコネクターを保存する環発生用コネクターのテーブルおよび長短様々のコネクターを保存する官能基結合用コネクターのテーブルの２種を用意することが好ましい。例えば、環発生用コネクターは１〜３原子からなり、官能基結合用コネクターは１〜１０原子からなる。
【００５２】
次に、Ｃ１１において、結合最小原子数を算出する。結合最小原子数とは、官能基間をＳＰ³ の炭素原子で鎖状につなげていった場合、つなげることができる最小の炭素原子数をいう。
次いで、Ｃ１２において、許容結合原子数を指定する。以下の操作において、結合最小原子数より許容結合原子数の数だけ原子数の多いコネクターを選択し、官能基をつなげることとなる。許容結合原子数は、通常、２〜６個が好ましい。
【００５３】
Ｃ１３において、構成する原子数が結合最小原子数と結合最小原子数＋許容結合原子数の間にあるコネクターの１つを選択する。
Ｃ１４において、コネクター内の回転可能な結合をＣ９で設定した回転様式に従って回転させる。回転させる度に、以下のＣ１５〜Ｃ１９の操作を繰り返すことにより、コネクターの適切なコンホメーションを見出すことができる。ただし、ATOMcon1とATOMcon2の結合およびATOMcon3とATOMcon4の結合は回転させないこととする。
【００５４】
次に、Ｃ１５において、Ｄ値およびＴ値を算出する。Ｄ値は上記の式（１）から、Ｔ値は上記の式（２）から算出することができる。
但し、官能基側のATOMhb1 および ATOMhb3の原子座標としては、既知のリガンド分子が単独の場合には、Ｃ１で入力した原子座標を使用し、既知のリガンド分子が複数の場合は、分子重ね合わせを行った分子のうちの一つの分子中のATOMhb1 および ATOMhb3に対応するそれぞれの原子の位置を原子座標として使用する。あるいは、それぞれの分子中のATOMhb1 および ATOMhb3に対応するそれぞれの原子の重心の位置を原子座標として使用してもよい。他にも、同等の効果が期待される位置があれば、順次それらの原子座標に置換できるようにしてもよい。例えば、Ｃ１で複数の既知のリガンド分子の原子座標を入力した場合には、各々の既知のリガンド分子中のATOMhb1 および ATOMhb3の原子座標に順次置換するようにしてもよい。
【００５５】
次いで、Ｃ１６において、Ｃ１５で算出したＤ値およびＴ値がＣ７で設定した限界値よりも小さいか否かを判定する。このＤ値およびＴ値の両方が限界値よりも小さい場合には、以下のＣ１７のステップに進み、それ以外の場合には、Ｃ２１のステップへ飛ぶ。すなわち、Ｄ値およびＴ値の両方が限界値よりも小さい場合には、コネクターは官能基をつなぐのに適したコンホメーションをとっているとみなすのである。
【００５６】
次に、Ｃ１７において、コネクターを構成する原子の情報（原子の種類と原子座標、およびダミー原子の原子座標）を保存する。
次いで、Ｃ１８において、コネクターを官能基間に嵌め込む。具体的には、コネクター側のATOMcon1、ATOMcon2、ATOMcon3およびATOMcon4をそれぞれ官能基側のATOMhb1 、ATOMhb2 、ATOMhb3 およびATOMhb4 に重ね合わせるようにして、コネクターを構成する原子の座標を置き換えることにより、官能基間をコネクターでつなぐことができる。
【００５７】
Ｃ１９において、官能基に結合手が残っている場合には、Ｃ１１〜Ｃ１８と同様の操作を行って、結合手の残っている官能基と水素を付加することができる原子間を環発生用コネクターでつなぐことにより環構造を導入する。この操作を環発生用コネクターのテーブルの中に保存されているすべてのコネクターについて繰り返す。また、結合手の残っている官能基と水素を付加することができる原子のすべての組み合わせについて上記の操作を行う。
【００５８】
次いで、Ｃ２０において、Ｃ１１〜１８と同様の操作を行って、コネクター中の水素を付加することができる２個の原子間を環発生用コネクターでつなぐことにより環構造を導入する。この操作を環発生用コネクターのテーブルの中に保存されているすべてのコネクターについて繰り返す。また、コネクター中の水素を付加することができる原子の組み合わせのすべてについて上記の操作を行う。
【００５９】
このように、環構造を導入することにより、リガンド分子の構造は安定化し、また、リガンド分子のバラエティに富んだ構造が作成される。
上記のＣ１９および２０の操作は、１〜４回繰り返してもよい。
Ｃ２１において、官能基の原子座標を変えるか否かを選択する。官能基の原子座標を変えない場合には、以下のＣ２２のステップに進むが、変える場合には、Ｃ１５のステップに戻り、官能基の新たな原子座標を用いてＤ値およびＴ値を算出し、Ｃ１６〜Ｃ２１の操作を繰り返す。
【００６０】
次いで、Ｃ２２において、コネクター内の回転可能な結合をすべて回転させたか否かを判定する。回転可能な結合をすべて回転させた場合には、以下のＣ２３のステップに進むが、回転させていない場合には、Ｃ１４のステップに戻り他の回転可能な結合を回転させる。
Ｃ２３において、構成する原子数が結合最小原子数と結合最小原子数＋許容結合原子数の間にあるコネクターのすべてを選択したか否かを判定する。コネクターのすべてを選択した場合には、以下のＣ２４のステップに進むが、すべてを選択していない場合には、Ｃ１３のステップに戻り未選択のコネクターを選択する。
【００６１】
Ｃ２４において、コネクターを結合させる２個の原子の組み合わせのすべてを選択したか否かを判定する。コネクターを結合させる２個の原子の組み合わせのすべてを選択した場合には、以下のＣ２５のステップに進むが、すべてを選択していない場合には、Ｃ６のステップに戻り未選択の２個の原子の組み合わせを選択する。この判定により、Ｃ５で設定した２個の原子の組み合わせのすべてが選択されることになる。
【００６２】
次に、Ｃ２５において、各官能基間に嵌め込まれたコネクターを組み合わせることにより、新たなリガンド分子の骨格を形成する。
Ｃ２６において、類似構造を有するリガンド分子骨格を除去する。例えば、Ｃ２５で形成されたすべてのリガンド分子骨格の構造の最少自乗重ね合わせを行い、最少自乗残差が一定値以下のものを類似構造として除去することができる。上記のステップで形成された多数のリガンド分子骨格は本質的に同じ構造に属するものが異なる経路として登録されてしまうことが起こり得る。そこで、上記のステップで得られた最初のリガンド分子骨格をリガンド分子骨格の類似性を判断する基準となる鋳型とし、２番目以降に得られるリガンド分子骨格を最少自乗重ね合わせにより順次重ね合わせていき、最少自乗残差が一定値以下のものは除去し、一定値以上のものは鋳型として登録し、鋳型グループを作成する。この鋳型グループを構成するリガンド分子骨格は互いに非類似である。
【００６３】
次に、Ｃ２７において、上記の操作で作成された全リガンド分子骨格のうちのひとつを選択する。
Ｃ２８において、リガンド分子骨格中の原子価の満たされていない原子に、結合距離および角度を算出して水素原子を付加する。原子価の満たされていない原子が存在しない場合には、このステップは省略される。
【００６４】
次いで、Ｃ２９において、リガンド分子構造の制限付き構造最適化を行う。
上記の操作で作成した新しいリガンド分子は、官能基間にコネクターを嵌め込むことによって作成されたために、構造に多少の歪みが出てくる。そこで、官能基間の位置や方向が維持されるように制限付きの構造最適化を行うことか好ましい。制限付き構造最適化は、プログラムAMBER （S.J. Weiner, et al., (1984), J. Am. Chem. Soc., 106, 765-784）を用いたＭＭ計算により行うことができる。
【００６５】
その後、Ｃ３０において、リガンド分子の構造が受容体の内孔内におさまるか否かを判定する。この判定には、三次元格子点情報を用いることができる。リガンド分子を構成する各原子に最も近い三次元格子点がリガンド分子によって占有可能であるならば、リガンド分子の構造が受容体の内孔内におさまると判定することができる。
【００６６】
リガンド分子の構造が受容体の内孔からはみ出ていない場合には、以下のＣ３１のステップに進み、内孔からはみ出ている場合には、Ｃ３２のステップへ飛ぶ。
次いで、Ｃ３１において、リガンド分子構造を出力する。
Ｃ３２において、全リガンド分子骨格が選択されたか否かを判定する。全リガンド分子骨格が選択されていない場合には、Ｃ２７に戻り未選択のリガンド分子骨格を選択して、Ｃ２８〜Ｃ３２を操作を繰り返す。
【００６７】
さらに、Ｃ３３において、Ｃ４で抽出した官能基を同等の相互作用をする官能基と置換するか否かを選択する。官能基を置換する場合には、Ｃ４のステップに戻り新たな官能基について、Ｃ５〜Ｃ３３の操作を繰り返す。置換可能な官能基は、表１に示したとおりである。
なお、受容体とリガンドの複合体の構造が解明されている場合には、上記の実施例のＲ１〜Ｒ１５、およびＣ１〜Ｃ４のステップに代えて、既知の受容体の構造に基づいて、受容体の内孔内に作成した三次元格子点の情報を算出し、かつ、受容体と直接に相互作用しているリガンド分子の官能基を抽出し、その官能基の構造座標を設定する以外は、実施例と同じ処理ステップを行うことによりリガンドの分子構造を構築することが可能である。
【００６８】
また、受容体単独の立体構造が解明され、薬物結合部位の環境がわかっている場合には、上記の実施例のＲ１〜Ｒ１５、およびＣ１〜Ｃ４のステップに代えて、受容体の内孔内に作成した三次元格子点の情報を算出し、かつ、受容体に存在する官能基と相互作用できる２個以上の官能基の種類と座標をユーザが設定する以外は、実施例と同じ処理ステップを行なうことにより、リガンド分子の分子構造を構築することが可能である。
【００６９】
〔例〕
本発明の方法を、ＴＰＡ、テレオシジン(teleocidin)Ｂ−４、アプリシアトキシン(aplysiatoxin)およびタプシガルギン(thapsigargin)からなる発癌プロモーターの系に適用してみた。これらの発癌プロモーターは化学構造が大きく異なるにもかかわらず、同一受容体に結合すると考えられている。本発明の方法に従って、上記の４種の発癌プロモーターの構造をもとにして官能基を抽出し、抽出した官能基をコネクターでつないでいくことにより、これらの発癌プロモーター作用を有する既知の化合物の分子構造が構築されれば、本発明の方法の有効性が実証される。
【００７０】
ＴＰＡ、テレオシジン、アプリシアトキシンおよびタプシガルギンの化学構造を図６に示す。
まず、ＴＰＡ、テレオシジン、アプリシアトキシンおよびタプシガルギンの各分子の重ね合わせをＲＥＣＥＰＳを用いて行い、得られた重ね合わせ構造を基に受容体モデルを構築した。ＴＰＡ、テレオシジン、アプリシアトキシンおよびタプシガルギンの原子座標としては、ケンブリッジ結晶構造データベース（ＣＣＳＤ）から得られる値を用いた。三次元格子点の間隔は0.５オングストロームとした。
【００７１】
また、ＲＥＣＥＰＳのＲ５のステップにおいては、自動重ね合わせを行うこととした。Ｒ１０のステップにおいて、分子形状、静電ポテンシャル、疎水性部分の一致のよい重ね合わせ構造を選択して重ね合わせモデルを最適化した。さらに、Ｒ１３のステップで、水素結合サイトとしてＯＨ基（水素供与性官能基）とＣ＝Ｏ基（水素受容性官能基）を選択した。図７に、ＲＥＣＥＰＳを用いた重ね合わせにより得られた各分子の立体構造を示す。また、ＲＥＣＥＰＳにより構築された受容体モデルを図８に示す。線図で示されたカゴで囲まれた空間が、リガンド分子に許容されると仮定された空間（受容体の内孔）を示す。図８には、ダミー原子を付加したリガンド分子の官能基（Ｃ＝Ｏ基とＯＨ基）も示されている。
【００７２】
次いで、ＣＯＮＮＥＣＴＯＲにより、リガンドの分子構造を構築した。Ｃ４のステップにおいては、ＯＨ基（水素供与性官能基）とＣ＝Ｏ基（水素受容性官能基）を抽出した。Ｃ７のステップにおいて、Ｄ値の限界値を３オングストロームに、Ｔ値の限界値を２０°に設定した。Ｃ８のステップでは、フラグメント集団として、４価の炭素、ベンゼン環、エチレン、アミド（−ＮＨＣＯ−）、シクロヘキサンを用意した。Ｃ９のステップにおいて、コネクター内の回転可能な単結合を０〜３６０°の範囲で３０°ずつ回転させるように結合の回転様式を設定した。また、Ｃ１５のステップにおいて、官能基側のATOMhb1 およびATOMhb3 の原子座標としては、分子重ね合わせを行った分子のうちＴＰＡ分子中のATOMhb1 およびATOMhb3 に対応するそれぞれの原子の位置を原子座標として使用した。
【００７３】
なお、Ｃ２１のステップで官能基の座標を変えず、また、Ｃ３３のステップで同等の相互作用をする官能基への置換を行わなかった。
その結果、約２００個の分子構造が得られた。その１部の分子構造を図９に示す。これらの分子構造には、ＴＰＡ(Hecker, 1. Carcinogenesis: A Comprehensive Survey, Mechanisms of Tumor Promotion and Carcinogenesis, edited by T.J. Slaga, A. Sivak and R.R. Boutwell, (1978) p.11-78) と同じ骨格を有する分子構造Ａ、大野、遠藤ら（M. Ohno, Y. Endo, M. Hirano, A. Itai, K. Shudo (1993) Tetrahedron Letters, in press)により発癌プロモーター作用が確認されている８員環ベンゾラクタムと類似の骨格を有する分子構造Ｂ、遠藤ら（Oxygenated Cholesterol as Ligands for Cytosolic-Nuclear Tumor Promoter: Yakkasterol, Y. Endo, Y. Hashimoto, H. Fukasawa, K. Shudo, B.B.R.C, 194, 1529-1935, (1993)) により発癌プロモーターの受容体との結合が確認されているヤッカステロールと類似の骨格を有する分子構造Ｃが含まれている。従って、その他の化合物およびその誘導体がＴＰＡ等と同等またはそれ以上の発癌プロモーター作用を示す可能性は極めて高い。
【図面の簡単な説明】
【図１】図１は、本発明の方法の一例における主要なステップを模式的に示したものである。
【図２】図２は、ＣＯＮＮＥＣＴＯＲを用いた本発明の方法の一実施態様のフローチャートの一部を示す。
【図３】図３は、ＣＯＮＮＥＣＴＯＲを用いた本発明の方法の一実施態様のフローチャートの一部を示す。
【図４】図４は、ＣＯＮＮＥＣＴＯＲを用いた本発明の方法の一実施態様のフローチャートの一部を示す。
【図５】図５は、ＲＥＣＥＰＳを用いた分子重ね合わせ法とレセプタマッピングのフローチャートの一部を示す。
【図６】図６は、テレオシジンＢ−４、ＴＰＡ、アプリシアトキシンおよびタプシガルギンの化学構造を示す。
【図７】図７は、ＲＥＣＥＰＳによる重ね合わせにより得られたテレオシジンＢ−４、ＴＰＡ、アプリシアトキシンおよびタプシガルギンの立体構造を示す。
【図８】図８は、三次元格子点を用いて表現した発癌プロモーターの受容体モデルを示す。
【図９】図９は、本発明の方法により構築された発癌プロモーターの受容体のリガンドの分子構造の一部を示す。[0001]
[Industrial application fields]
The present invention relates to a method for constructing a molecular structure of a ligand compound having physiological activity. More specifically, the present invention relates to a molecular structure of a novel ligand compound that can be used for designing the molecular structure of a pharmaceutical, agricultural chemical, or other compound having physiological activity. The present invention relates to a method for automatic generation.
[0002]
[Prior art]
There are conceptually two steps in drug design. The lead creation process finds a compound (referred to as lead) as a starting point for the subsequent drug development, and the lead optimization process finds a compound with higher activity and less toxicity and side effects starting from the lead. Lead is a compound that is known to have a certain physiological activity and is also referred to as a lead compound. Most of the drug design generally performed is this second step, and is often performed by synthesizing a series of substituent derivatives from known active compounds and comparing their activities. It is extremely difficult to perform the first process artificially, discovery from natural substances, discovery from known side effects of drugs, modification of structures from biologically active substances (neurotransmitters, hormones, vitamins, etc.) Alternatively, as soon as the available compounds are handled, there is a great deal of reliance on chance, luck, and trial and error, such as the discovery by random screening on a prepared evaluation system. Starting from an appropriate lead will determine the success or failure of subsequent drug development, and there are many diseases in which a good drug cannot be made because there is no appropriate lead. It is widely desired. What kind of pharmacological activity a certain compound shows depends on which receptor (biopolymer) binds, but one compound does not always bind to only one receptor. Therefore, the strength of the main activity and the side activity (side effects) may be reversed or the activity may disappear due to a slight difference in structure. Even if the main activity does not change, the behavior in the body such as ease of metabolism and interaction with the membrane changes greatly, often affecting the properties as a drug. In addition, drugs from new leads with different molecular skeletons are expected to behave quite differently, although they have the same main activity.
[0003]
In order for a drug to exhibit physiological activity, it must reach a certain site of the receptor in the body and specifically interact with the receptor (usually a strong non-covalent association). In recent years, the technology for receptor isolation and purification has progressed, and the number of three-dimensional structures (three-dimensional coordinates of constituent atoms) of enzymes and nucleic acids has been clarified by X-ray crystallography. In particular, analysis of complex with ligand (inhibitor in the case of enzyme) shows strong van der Waals interaction between ligand molecule showing strong binding and receptor due to good fit of molecular surface. In addition, it has been found that hydrophilic interactions such as hydrogen bonding and electrostatic interaction contribute to the stabilization of the complex. In spite of the fact that the apparent structure is completely different from the active substances and enzyme substrates that are originally working in the living body, the binding to the same receptor has been confirmed, and many of them have the same action or inhibitors. This also indicates that what is required for strong binding between the receptor and the drug is the three-dimensional shape of both molecules and the complementarity of these properties. Therefore, if humans can design a molecular structure with a three-dimensional structure that satisfies these requirements, lead creation should be artificially necessary. However, since humans are easily trapped by the structure of known active compounds and are not good at three-dimensional insight, it is difficult to create leads even when the three-dimensional structure of the receptor is elucidated by crystal analysis. Therefore, an approach in which such a structure is presented to a computer and then selected by humans is considered to be effective in finding new active structures and using them as leads. Computers are not preoccupied with preconceptions, but cover all possibilities and give results based on quantitative evidence.
[0004]
There are two possible approaches to finding a structure that meets the requirements of the lead. There are a method for searching a structure that meets a condition from a structure database of known compounds, and a method for newly constructing a structure that meets such a condition.
When the three-dimensional structure of the receptor is available, the following reports are available for both approaches.
[0005]
DesJarlais, Kuntz et al. Expressed the shape of the inner pore of the drug binding site with several spheres that fit well there, and published a crystal database (published by the University of Cambridge, UK: reported to over 100 chemical journals around the world) Announced a method (DOCK) that expresses each compound in a registered sphere as a few spheres containing it and determines the similarity to the receptor sphere collection (RL DesJarlais, RP Sheridan , GLSeibel, JSDixon, ID Kuntz and R. Venkataraghavan, J. Med. Chem. 31, 722-729 (1988); RL DesJarlais, GL Seibel, ID Kuntz, PS Furth and JC Alvarez, Proc. Natl. Acad. Sci. USA, 87, 6644-6648 (1990)). As a result of applying this method to the AIDS virus protease (HIV protease: registered in the Protein Data Bank), the crystal structure of a compound called haloperidol, known as a beta-blocker and antipsychotic, is often used as the drug-binding site of this enzyme. It was selected as a fit, synthesized with minor structural modifications, and tested for activity, which proved its effectiveness in treating AIDS by inhibiting this enzyme. However, this method only determines the shape of the molecule, does not consider hydrogen bonds or other interactions, and considers only the crystal structure of the molecular conformation, so leakage of promising candidate compounds and inefficiency are problems. Yes.
[0006]
In addition, Lewis et al. Assume a diamond lattice having a side length of a carbon-carbon single bond at the drug binding site, and a method for newly constructing a molecular structure by appropriately installing carbon atoms on the lattice points. (RA Lewis, J. Comp.-Aided Molecular Design, 4, 205-210 (1990)). But this way sp ^Three Atoms other than hybrids cannot be set up, and specific interactions with receptors are not considered at all.
[0007]
Nishibata and Itai et al. Announced a method of assembling a structure that fits well into the drug binding site of the receptor and has a stable interaction, and a program LEGEND that implements this method (Y. Nishibata and A. Itai, Tetrahedron, 47 , 8986-8990 (1991)). In this method, atoms or atomic groups are added one by one based on random numbers and force fields so that van der Waals interaction, hydrogen bonding, and electrostatic interaction with the acceptor are advantageous. . It is excellent in suggesting a large number of possible structures, but the effectiveness of the application has been proven, but no actual success has been confirmed.
[0008]
Bohm has developed a method for constructing a new structure by inserting several appropriate fragment structures prepared in the program into the drug binding site of the receptor and linking them well (H. -J. Bohm, J. of Comp.-Aided Molecular Design, 6, 61-78 (1992)). However, no results have yet been proven to prove the effectiveness of this method.
[0009]
If the three-dimensional structure of the receptor is not known, the above two approaches cannot be used, so another approach is necessary.
The number of biopolymers whose steric structures are currently announced is about a thousand, including those derived from bacteria. The structure of human receptors, which are said to exist in the body, is determined to be one by one. Is expected to take a long time. Therefore, an approach for constructing a structure of an active ligand molecule when there is no information on the three-dimensional structure of the receptor is currently particularly required.
[0010]
When there is no information on the three-dimensional structure of the receptor, the most effective method for finding the relationship between structure and activity from the side of ligand molecules such as drugs, bioactive substances, and bioactive substances is molecular superposition. . Multiple compounds that bind to the same site on the same receptor and express similar physiological effects (though not overlapping when molecular structures are similar) have common features in shape and intermolecular interactions The common features can be extracted by superimposing these molecules three-dimensionally. In addition, the determination of the molecular conformation (active conformation or active conformation) when a molecule that has a single bond that can rotate in the molecule binds to the receptor and expresses the activity is also sometimes performed. Made possible by. In addition, estimating the environment and size of the drug-binding site of the receptor from the ligand side is called receptor mapping, and that estimated is called a receptor model. This receptor mapping can be effectively performed based on the superposition of molecules.
[0011]
There are various methods for superimposing molecules, but the results of superposition differ greatly depending on the method. Different structure-activity relationship models and receptor models (estimated functional groups, shape and properties of receptor drug binding sites, Environmental estimation). In order to reach the correct model as much as possible, it is essential to use a logical superposition method.
[0012]
As a method of superimposing molecules, which has been generally performed so far, three or more corresponding atom pairs are designated based on the similarity of the positions of the molecular skeleton and heteroatoms, and the positions of the two molecules are determined between the two molecules. In order to achieve the best match, it has been superposed by least squares calculation or appropriately superposed by visual judgment on a computer graphics screen.
[0013]
However, Itai and Kato et al. Believe that the atomic positions (molecular skeletons) do not need to match between the molecules of multiple drugs or bioactive substances that bind to the same receptor. The molecular superposition method (RECEPS) has been devised and published so that the shape, hydrogen bond, electrostatic properties and other physical and chemical properties of each other coincide with each other. (Y. Kato, A. Itai and Y. Iitaka, Tetrahedron, 43, 5229-5236 (1987); A. Itai, Y. Kato, N. Tomioka, Y. Iitaka, Y. Endo, M. Hasegawa, K. Shudo, H. Fujiki and S. Sakai, Proc. Natl. Acad. Sci. USA, 85, 3688-3692 (1988); Y. Kato, A. Inoue, M. Yamada, N. Tomioka and A. Itai, J Comp.-Aided Molecular Design, 6 in press (1992); A. Itai, N. Tomioka, Y. Kato, Y. Nishibata and S. Saito, in 'Medicinal Chemistry in 21st Century', edited by CG Wermuth, Blackwell Scientific Publishings, London, (1992) in press; A. Itai, N. Tomioka and Y. Kato, in 'QSAR: New developments and Applications', edited by T. Fujita, Elsevier Science Publishers, Amsterdam, (1992), in Using this RECEPS, it is possible to obtain a good overlay while changing the conformation by moving and rotating the molecules interactively on the computer graphics screen using the coincidence between these molecules as an index. It became. This made it possible for the first time to overlap molecules with completely different chemical structures.
[0014]
Then, in this RECEPS, in a system in which hydrogen bonding plays an important role, the computer automatically covers a superposition structure that matches the positions of hydrogen bonding functional groups expected from a plurality of molecules on the acceptor side, Added a function that ranks and shows. This improvement has resulted in objective, correct overlay results that cover a variety of possibilities derived from overlay methods and conformations.
[0015]
Using the molecular superposition method as described above, it is possible to extract features common to multiple ligand molecules with completely different structures when the three-dimensional structure of the receptor is not known. The information obtained by the molecular superposition method can only explain the relationship between structure and activity retrospectively, and a method for designing a new ligand molecule based on this information and creating a lead has not yet been established. .
[0016]
[Problems to be solved by the invention]
Accordingly, an object of the present invention is to provide a method for constructing a structure of a novel ligand molecule having activity based on the structure of a known active ligand molecule.
Another object of the present invention is to provide a method for constructing a structure of a novel ligand molecule having activity based on the molecular structure of a known receptor.
[0017]
[Means for Solving the Problems]
As a result of diligent efforts, the inventors of the present invention have three-dimensionally arranged hydrogen bonding functional groups that are considered necessary for binding to the receptor, and while maintaining this position and direction, these hydrogen bonding functional groups are We have succeeded in solving the above problems by developing a method for automatically constructing a molecular structure of a novel ligand by connecting a group of atoms called a connector.
[0018]
That is, the present invention
A method for constructing a molecular structure of a ligand having physiological activity,
(i) a first step of disposing two or more functional groups presumed to be essential for the physiological activity of the ligand and designating the position and direction of the functional groups;
(ii) a second step of setting a combination of two functional groups to be connected among the functional groups arranged in the first step;
(iii) a third step of creating a table of a series of atomic groups for constructing the molecular structure of the ligand;
(iv) From the table of a series of atomic groups created in the third step, the one connecting the two functional groups set to be connected in the second step is selected, and a bonding route is created between the two functional groups. A fourth step of forming a skeleton of the ligand molecule by, and
(v) Fifth step of determining whether the molecular structure of the constructed ligand is compatible with the inner pore of the receptor
The above method is provided.
[0019]
The present invention also provides a method for constructing a molecular structure of a ligand having physiological activity,
(i) a first step of disposing two or more functional groups presumed to be essential for the physiological activity of the ligand and designating the position and direction of the functional groups;
(ii) a second step of setting a combination of two functional groups to be connected among the functional groups arranged in the first step;
(iii) a third step of creating a table of a series of atomic groups for constructing the molecular structure of the ligand;
(iv) From the table of a series of atomic groups created in the third step, the one connecting the two functional groups set to be connected in the second step is selected, and a bonding route is created between the two functional groups. The fourth step to
(v) a fifth step of forming a ligand molecule skeleton by combining the binding pathways created by repeating the fourth step for all combinations of two functional groups set to be connected in the third step; and
(vi) Sixth step of determining whether the molecular structure of the constructed ligand is compatible with the inner pore of the receptor
The above method is provided.
[0020]
The method for constructing the molecular structure of a ligand having physiological activity of the present invention may include a step of introducing a ring structure after forming a binding pathway between two functional groups.
In addition, the position and direction of the functional group specified in the first step are sequentially specified in the position and direction estimated to have the same effect, and the second to fifth to sixth steps are repeated. May be. Further, the functional groups arranged in the first step may be sequentially replaced with other functional groups having equivalent physicochemical properties, and the second to fifth steps to the sixth step may be repeated.
[0021]
Further, in addition to the first step to the fifth step to the sixth step, it is determined whether or not there is an atom whose valence is not satisfied in the molecular skeleton of the ligand. May include a step of adding a hydrogen atom to the atom and a step of optimizing the molecular structure of the ligand.
If there is a known ligand molecule but the receptor structure is not known, the internal pores of the ligand receptor can be obtained by obtaining three-dimensional lattice point information based on the molecular structure of one or more known ligands. After the model is constructed, the molecular structure of the ligand may be constructed as described above.
[0022]
According to the method of the present invention, when a known ligand molecule is present but the structure of the receptor is not known, or when the structure of the complex of the receptor and the ligand molecule is elucidated, the structure of the receptor alone is analyzed. The structure of the drug binding site has been elucidated by this method, but even when there is no known ligand molecule, a novel ligand molecular structure can be constructed.
[0023]
In the present invention, the term “receptor” refers to any biopolymer that can be a target of a drug, and includes proteins, nucleic acids, polysaccharides, and complexes thereof.
Ligand means a low molecular weight compound (about 1,000 or less) that strongly binds (interacts) with a specific site of a specific receptor, such as neurotransmitters, hormones, vitamins and other biologically active substances, natural And synthetic physiologically active substances, drugs, enzyme substrates and the like.
[0024]
The functional group means an atom considered to have a specific function or a group of atoms (atomic group) always acting as a group in chemical reaction or molecular recognition. For example, an amino group, a nitro group, a benzene ring and the like can be exemplified.
The direction of the functional group means a direction capable of interacting with the receptor. If the functional group is a hydrogen bond, the direction of the lone electron pair or hydrogen is used. If the aromatic ring is a benzene ring, the direction of the π electron perpendicular to the ring is the functional group direction.
[0025]
Three-dimensional lattice point information expresses the receptor model spatially and position-dependently, so that the lattice spacing is about 0.2 to 0.5 angstroms, and the environment information of the receptor at that position on each lattice point. It shall mean data that has been calculated and saved. The three-dimensional lattice point information includes, for example, an address indicating a spatial position for each lattice point, occupation information by each ligand molecule (up to six molecules), electrostatic potential, and whether a hydrogen bonding functional group is expected. If so, it includes data such as whether the property is hydrogen donating or accepting or whether the aromatic property is a problem. Occupancy information by the ligand molecule is a lattice point that indicates whether or not the three-dimensional lattice point is in the region occupied by the ligand molecule, that is, in the region included in the van der Waals radius of the atoms constituting the ligand molecule. Information for each. This three-dimensional lattice point information can be used instead of the three-dimensional structure of the receptor, and the molecular structure of the ligand can be constructed while judging whether or not it matches.
[0026]
【The invention's effect】
By the method of the present invention, a novel ligand compound having physiological activity can be predicted with high probability.
In addition, the method of the present invention enables high-speed molecular design of a novel ligand compound having physiological activity.
[0027]
【Example】
FIG. 1 schematically illustrates the main steps in an example of the method of the present invention. In an example of the method of the present invention, first, receptor mapping is performed based on the molecular structure of a known ligand, for example, by the RECEPS technique, three-dimensional lattice point information is obtained, and then a receptor model is constructed. By using the CONNECTOR method, a functional group of a known ligand molecule that is important for interaction with a receptor is extracted, the position and direction of the functional group are designated, and the position and direction of the functional group are preserved and new. The molecular structure of a simple ligand can be constructed. The C═O and OH groups in FIG. 1 are functional groups extracted from the molecular structure of known ligands that are important for the interaction with the receptor.
[0028]
2, 3 and 4 show a flow chart of one embodiment of the method of the present invention using the CONNECTOR approach. Details will be described later.
FIG. 5 shows a flowchart of the molecular superposition method and receptor mapping using the RECEPS method.
FIG. 6 shows the chemical structures of teleocidin B-4, TPA, aprisia toxin and thapsigargin.
[0029]
FIG. 7 shows the three-dimensional structure of teleocidin B-4, TPA, aprisia toxin, and thapsigargin obtained by superposition by RECEPS.
FIG. 8 shows an oncogenic promoter receptor model expressed using three-dimensional lattice points and functional groups of ligand molecules to which dummy atoms are added. A space surrounded by a cage represented by a diagram shows the inner hole of the receptor.
[0030]
FIG. 9 shows an example of the molecular structure of a ligand of a receptor for an oncopromoter constructed by the method of the present invention.
An embodiment of the present invention will be described with reference to the flowcharts of FIGS. FIG. 5 shows a flow from inputting a known ligand molecule coordinate by the RECEPS method to obtaining three-dimensional lattice point information, and FIGS. 2 to 4 show the atomic coordinates of the ligand molecule and the CONNECTOR method. The flow from inputting the three-dimensional lattice point information obtained by the RECEPS method until the molecular structure of a new ligand is output is shown. 2 to 5, R and C indicate each step.
[0031]
First, in R1 of FIG. 5, atomic coordinates of one kind of ligand molecule known to have physiological activity or two or more kinds of ligand molecules known to have similar physiological activity are input. Hereinafter, the ligand molecule input to the atomic coordinates is referred to as a known ligand molecule.
As the structure of the known ligand molecule, a structure determined by crystal analysis, a similar structure in a crystal database, or a structure obtained by molecular modeling using energy calculation can be used.
[0032]
Next, in R2, for each known ligand molecule, the charge of each atom is calculated by molecular orbital calculation, and a type number is assigned to each hydrogen-bonding functional group for each functional group.
Here, the hydrogen bonding functional group refers to a functional group containing hydrogen for hydrogen bonding or a hetero atom having a pair of isolated atoms.
[0033]
Next, in R3, it is determined whether or not there is one kind of known ligand molecule. When there is one kind of known ligand molecule, the step of R11 is performed without performing the steps of R4 to R10 described below.
When there are two or more types of known ligand molecules in R3, two types of molecules are selected from the known ligand molecules in R4, and one is a template molecule and the other is a trial molecule.
[0034]
Next, in R5, it is selected whether to perform automatic superposition of known ligand molecules or interactive superposition.
Here, the superposition of molecules refers to placing three-dimensional structures of multiple active molecules in corresponding orientations in the same space in order to three-dimensionally explore and compare the homology and differences in molecular shape and properties. Say. The question is what kind of properties and physical quantities to focus on (shape of surface of molecule, position of contained atoms, physical and chemical properties, etc.), and how to evaluate the goodness of overlap. For example, in RECEPS, superposition can be performed with an emphasis on matching the properties of molecules such as hydrogen bonds, ionic bonds, and hydrophobic interactions for interacting with receptors. Molecular superposition may be performed using a molecular model, but it is preferable to use computer graphics because it is easy, can be quantified, and coordinates can be stored as they are. In addition, the automatic superposition of molecules refers to a method in which a computer automatically finds a superposition method while covering the positional relationship and molecular conformation of superposition. For example, in RECEPS, a special least square method calculation is performed so that hydrogen bonds on the acceptor side coincide with each other on the assumption that the same acceptor binds to the same site. In addition, the interactive superimposition of molecules refers to superimposing molecules on a computer graphics display while rotating, translating, and coupling and rotating the molecules while referring to the index of overlap.
[0035]
In the R5 step, it is generally preferable to select automatic superposition. However, when there are few hydrogen bonding sites in the molecule and it is difficult to use the automatic superposition method, interactive superposition may be selected. .
If automatic superposition of known ligand molecules is selected in R5, then automatic superposition using dummy atoms is performed in R6. A dummy atom is based on the position of a group (eg, a hydrogen donating group) involved in the interaction of one molecule when there are two molecules (eg, a receptor and a ligand) that form a hydrogen bond. , A virtual atom placed at a position presumed to be suitable as the position of a group (for example, a hydrogen accepting group) involved in the interaction of the other molecule. For example, in a receptor-ligand system in which two or more hydrogen bonds are presumed to be involved in ligand activity, they are arranged on the receptor side based on the position and nature of the hydrogen-bonding functional group of the template molecule. Based on the position and nature of the hydrogen bonding functional group of the trial molecule, the positional relationship between the template molecule and the trial molecule, and the template molecule A model in which the template molecule and the trial molecule are superimposed can be created while changing the conformation of the trial molecule. All known ligand molecules are superposed by repeating automatic superposition with a template molecule instead of the trial molecule (Y. Kato, A. Inoue, M. Yamada, N. Tomioka, and A. Itai, Computer-Aided Molecular Design 6, pp.475-486 (1992)).
[0036]
If interactive superposition is selected in R5, next, in R7, an area for creating a three-dimensional lattice point is set for the template molecule.
Next, in R8, three-dimensional lattice point information (shape, hydrogen bondability, electrostatic potential, aromaticity, etc.) for the template molecule is calculated and stored.
Next, in R9, the reference molecule is interactively superimposed on the template molecule while referring to an index for coincidence of physicochemical properties calculated in real time using the above three-dimensional lattice point information. All known ligand molecules are superposed by repeating superposition with the template molecule instead of the trial molecule.
[0037]
Next, in R10, the automatic superposition model of the template molecule and the trial molecule obtained in R6 or the interactive superposition model of the template molecule and the trial molecule obtained in R9 is optimized.
All superposition models with good superposition indices obtained in R6 or R9 are re-evaluated using indices (hydrogen bondability, electrostatic potential, shape, aromaticity, etc.) using three-dimensional lattice points, and Get a promising overlay model.
[0038]
Alternatively, using the superposition result obtained in R6 or R9 as the initial structure, the molecular structure can be manipulated interactively to improve the superposition model, or the superposition model can be created by the Simplex method using three-dimensional lattice point data. To obtain an appropriate overlay model.
Next, in R11, an area for creating a three-dimensional lattice point based on the superposed structure of the known ligand molecules obtained in steps R2 to R10 or the structure of a single known ligand molecule input in R1. Set.
[0039]
Next, in R12, three-dimensional lattice point information for all molecules is calculated within the region set in R11.
Receptor mapping can be performed based on this three-dimensional lattice point information. Here, receptor mapping means estimating and expressing the environment of the drug binding site of the receptor from the information of the ligand molecule when there is no experimental information about the three-dimensional structure of the receptor. In RECEPS receptor mapping, the inner pores of a receptor model are used based on the structure of a single known ligand molecule or a plurality of known ligand molecules superimposed using three-dimensional lattice points inside and around the molecule. The size, shape, expected site of the hydrogen bonding functional group, electrostatic properties, aromaticity, etc. are expressed. The space occupied by three-dimensional lattice points occupied by one or more molecules is assumed to be the inner pore of the receptor. For the electrostatic potential in the outer region, an average value weighted by activity at each three-dimensional lattice point is adopted.
[0040]
Next, in R13, a hydrogen bonding site is set.
When there are multiple known ligand molecules, a hydrogen bonding functional group that is highly common between the stacked molecules is set as a hydrogen bonding site. The hydrogen bonding functional group that seems to be important is set as the hydrogen bonding site. Moreover, although it is not a hydrogen bondable functional group, such as a benzene ring, you may set together with said hydrogen bondable functional group the group considered to be essential for activity as a hydrogen bond site.
[0041]
Next, in R14, the three-dimensional lattice point information is output to a file.
Next, in R15, the atomic coordinates of a known superposition model of ligand molecules are output.
Next, in C1 of FIG. 2, the atomic coordinates of the known single ligand molecule or the atomic coordinates of the overlay model of the known ligand molecule output in R15 are input.
[0042]
Next, in C2, the charge of each atom is calculated in advance by molecular orbital calculation for each ligand molecule, and the type number for each functional group is given to the hydrogen bonding functional group at the user's discretion. deep.
Next, in C3, the three-dimensional lattice point information calculated in R12 is input.
Next, in C4, a functional group presumed to be essential for activity is extracted from the above-mentioned known ligand molecule or ligand molecule group that interacts with the receptor.
[0043]
Furthermore, a new ligand molecule structure can be constructed by preparing a plurality of functional groups having the same interaction as fragments of these functional groups and sequentially replacing them. Examples of the substitutable functional groups are shown in Table 1 below.
[0044]
[Table 1]

Next, in C5, all combinations of two atoms to which the connector is bonded are set out of the atoms constituting the functional group extracted in the above-described C4 step. For example, when> NH and -OH are selected in the step of C4, an N atom of> NH and an O atom of -OH are selected. Here, the connector refers to a series of atomic groups connecting the functional groups extracted in the step C4.
[0045]
Next, in C6, one is selected from the combination of the two atoms set in C5. The two atoms selected here will be referred to as ATOMhb1 and ATOMhb3. The dummy atom added to ATOMhb1 is called ATOMhb2 and the dummy atom added to ATOMhb3 is called ATOMhb4. Here, the dummy atom means a virtual atom to be added for determining the suitability of the connector. The positions of the dummy atoms ATOMhb2 and ATOMhb4 are inevitably determined by the position and direction of the extracted functional group. Hereinafter, ATOMhb1 to ATOMhb4 are referred to as functional group-side atoms ATOMhb1 to ATOMhb4.
[0046]
Next, in C7, a limit value for the D value and a limit value for the T value are set.
The D value is calculated from the following formula (1). The distance relationship between the four atoms ATOMhb1, ATOMhb2, ATOMhb3 and ATOMhb4 on the functional group side and the distance relationship between the four atoms ATOMcon1, ATOMcon2, ATOMcon3 and ATOMcon4 on the connector side This shows the similarity. Here, ATOMcon2 and ATOMcon4 represent atoms at both ends of the connector, and ATOMcon1 and ATOMcon3 represent dummy atoms added to both ends of the connector.
[0047]
D value = (D1) ² + (D2) ² + (D3) ² + (D4) ² Formula (1)
(Where D1 is the (distance between ATOMhb1 and ATOMhb3) and (the distance between ATOMcon1 and ATOMcon3), D2 is the (distance between ATOMhb2 and ATOMhb4) and (the distance between ATOMcon2 and ATOMcon4), and D3 is (the distance between ATOMhb1 and ATOMhb1). D4 represents the difference between (distance between ATOMhb4) and (distance between ATOMcon1 and ATOMcon4), and D4 (distance between ATOMhb2 and ATOMhb3) and (distance between ATOMcon2 and ATOMcon3).
The smaller the D value, the more similar the distance relationship between the four atoms.
[0048]
The T value is calculated from the following formula (2). The twist angle of the four atoms ATOMhb2, ATOMhb1, ATOMhb3 and ATOMhb4 on the functional group side and the twist angle of the four atoms ATOMcon2, ATOMcon1, ATOMcon3 and ATOMcon4 on the connector side The difference is shown.
T value = T _hb -T _con Formula (2)
(Where T _hb Gives the torsion angles of ATOMhb2, ATOMhb1, ATOMhb3 and ATOMhb4, T _con Represents the torsion angles of ATOMcon2, ATOMcon1, ATOMcon3 and ATOMcon4. )
The smaller the T value, the more similar the positional relationship between the four atoms.
[0049]
The limit value of the D value and the limit value of the T value can be freely selected by the user, but if the D value is in the range of 2 to 5 angstroms and the T value is in the range of 10 to 20 °, it is accurate. It is also preferable because an appropriate connector can be selected efficiently.
Next, in C8, a fragment group is prepared. A fragment refers to a partial structure for creating a connector. As a fragment,
Tetravalent carbon (SP ^Three Atoms such as C) and functional groups such as benzene ring, ethylene, carbonyl, cyclohexane and the like.
[0050]
Then, in C9, the rotational mode of the rotatable coupling in the connector is set. In most cases, there are rotatable bonds within the connector, and by systematically rotating them, it is possible to find a conformation of the connector that fits well between functional groups. For example, in the case of a rotatable single bond, it is possible to set a rotation mode that systematically changes the torsion angle in steps of 20 ° to 60 °, and when the rotatable bond is in the ring structure, it is possible. Various ring structures can be set so that they can be input sequentially from a file.
[0051]
Next, in C10, connectors of various lengths are created by combining all fragments of the fragment group prepared in C8, and a table storing these is prepared. It is preferable to prepare two types of tables: a ring-generating connector table that stores short connectors and a functional group-bonding connector table that stores various long and short connectors. For example, the ring generating connector comprises 1 to 3 atoms, and the functional group bonding connector comprises 1 to 10 atoms.
[0052]
Next, in C11, the minimum number of bonds is calculated. The minimum number of bonds is SP between functional groups. ^Three This is the minimum number of carbon atoms that can be connected if they are connected in a chain.
Next, in C12, the allowable number of bonding atoms is designated. In the following operation, a connector having a larger number of atoms than the minimum number of bonding atoms by the number of allowable bonding atoms is selected, and the functional group is connected. The number of allowable bond atoms is usually preferably 2-6.
[0053]
In C13, one of the connectors in which the number of constituent atoms is between the minimum number of bonds and the minimum number of bonds + allowable bond atoms is selected.
At C14, the rotatable coupling in the connector is rotated according to the rotation mode set at C9. The appropriate conformation of the connector can be found by repeating the following operations C15 to C19 each time it is rotated. However, the connection between ATOMcon1 and ATOMcon2 and the connection between ATOMcon3 and ATOMcon4 are not rotated.
[0054]
Next, in C15, a D value and a T value are calculated. The D value can be calculated from the above equation (1), and the T value can be calculated from the above equation (2).
However, as the atomic coordinates of ATOMhb1 and ATOMhb3 on the functional group side, if there is a single known ligand molecule, the atomic coordinates entered in C1 are used. If there are multiple known ligand molecules, molecular superposition is performed. The position of each atom corresponding to ATOMhb1 and ATOMhb3 in one of the molecules performed is used as atomic coordinates. Alternatively, the position of the center of gravity of each atom corresponding to ATOMhb1 and ATOMhb3 in each molecule may be used as atomic coordinates. In addition, if there is a position where an equivalent effect is expected, the atomic coordinates may be sequentially replaced. For example, when the atomic coordinates of a plurality of known ligand molecules are input in C1, the atomic coordinates of ATOMhb1 and ATOMhb3 in each known ligand molecule may be sequentially substituted.
[0055]
Next, at C16, it is determined whether the D value and T value calculated at C15 are smaller than the limit values set at C7. When both the D value and the T value are smaller than the limit value, the process proceeds to the following step C17, and otherwise, the process jumps to the step C21. That is, when both the D value and the T value are smaller than the limit values, the connector is considered to have a suitable conformation for connecting the functional groups.
[0056]
Next, in C17, information on the atoms constituting the connector (atom type and atomic coordinates, and atomic coordinates of dummy atoms) is stored.
The connector is then fitted between the functional groups at C18. Specifically, ATOMcon1, ATOMcon2, ATOMcon3, and ATOMcon4 on the connector side are overlapped with ATOMhb1, ATOMhb2, ATOMhb3, and ATOMhb4 on the functional group side, respectively, and the coordinates of the atoms that make up the connector are replaced. Can be connected with a connector.
[0057]
In C19, when a bond remains in the functional group, the same operation as in C11 to C18 is performed, and a ring generating connector is formed between the functional group remaining in the bond and an atom to which hydrogen can be added. Ring structure is introduced by connecting with This operation is repeated for all connectors stored in the ring generation connector table. In addition, the above-described operation is performed for all combinations of the functional group remaining in the bond and the atom to which hydrogen can be added.
[0058]
Next, in C20, a ring structure is introduced by connecting the two atoms capable of adding hydrogen in the connector with a ring-generating connector by performing the same operation as in C11-18. This operation is repeated for all connectors stored in the ring generation connector table. In addition, the above operation is performed for all combinations of atoms that can add hydrogen in the connector.
[0059]
Thus, by introducing a ring structure, the structure of the ligand molecule is stabilized, and a variety of structures of the ligand molecule are created.
The above operations of C19 and 20 may be repeated 1 to 4 times.
In C21, it is selected whether or not to change the atomic coordinates of the functional group. If the atomic coordinates of the functional group are not changed, the process proceeds to the following step C22. If changed, the process returns to the step C15 to calculate the D value and the T value using the new atomic coordinates of the functional group. , C16 to C21 are repeated.
[0060]
Next, in C22, it is determined whether or not all the rotatable couplings in the connector have been rotated. When all the rotatable couplings are rotated, the process proceeds to the following step C23, but when not rotated, the process returns to the step C14 to rotate other rotatable couplings.
In C23, it is determined whether or not all the connectors in which the number of constituent atoms is between the minimum number of bonds and the minimum number of bonds + allowable bond atoms are selected. If all of the connectors are selected, the process proceeds to the following step C24. If not all are selected, the process returns to the step C13 to select an unselected connector.
[0061]
In C24, it is determined whether or not all the combinations of two atoms to which the connector is bonded have been selected. When all the combinations of two atoms to which the connector is bonded are selected, the process proceeds to the following C25 step, but when not all are selected, the process returns to the C6 step and two unselected atoms. Select a combination. By this determination, all the combinations of two atoms set in C5 are selected.
[0062]
Next, in C25, a skeleton of a new ligand molecule is formed by combining connectors fitted between the functional groups.
In C26, the ligand molecular skeleton having a similar structure is removed. For example, the least squares superposition of the structures of all the ligand molecular skeletons formed of C25 can be performed, and those having a least square residual of a certain value or less can be removed as similar structures. In many ligand molecular skeletons formed in the above steps, those belonging to essentially the same structure may be registered as different routes. Therefore, the first ligand molecule skeleton obtained in the above step is used as a template as a reference for judging the similarity of the ligand molecule skeleton, and the second and subsequent ligand molecule skeletons are sequentially superposed by least square superposition. If the least-squares residual is less than a certain value, it is removed. The ligand molecular skeletons constituting this template group are dissimilar to each other.
[0063]
Next, in C27, one of all ligand molecular skeletons created by the above operation is selected.
In C28, a hydrogen atom is added to an atom whose valence is not satisfied in the ligand molecular skeleton by calculating a bond distance and an angle. This step is skipped if there are no uncharged atoms.
[0064]
Next, in C29, a limited structure optimization of the ligand molecular structure is performed.
Since the new ligand molecule created by the above operation is created by inserting a connector between functional groups, some distortion occurs in the structure. Therefore, it is preferable to perform limited structure optimization so that the positions and directions between the functional groups are maintained. Constrained structural optimization can be performed using the program AMBER (SJ Weiner, et al., (1984), J. Am. Chem. Soc., 106, 765-784).
[0065]
Thereafter, in C30, it is determined whether or not the structure of the ligand molecule is within the inner pore of the receptor. Three-dimensional lattice point information can be used for this determination. If the three-dimensional lattice point closest to each atom constituting the ligand molecule can be occupied by the ligand molecule, it can be determined that the structure of the ligand molecule falls within the inner pore of the receptor.
[0066]
When the structure of the ligand molecule does not protrude from the inner pore of the receptor, the process proceeds to the following step C31, and when it protrudes from the inner hole, the process jumps to the step C32.
Next, in C31, the ligand molecular structure is output.
In C32, it is determined whether or not all ligand molecular skeletons have been selected. When all the ligand molecule skeletons are not selected, the process returns to C27, selects an unselected ligand molecule skeleton, and repeats the operations from C28 to C32.
[0067]
Further, in C33, it is selected whether or not to replace the functional group extracted in C4 with a functional group having an equivalent interaction. When substituting a functional group, the process returns to the C4 step and the operation of C5 to C33 is repeated for the new functional group. Substitutable functional groups are as shown in Table 1.
In addition, when the structure of the complex of the receptor and the ligand has been elucidated, instead of the steps R1 to R15 and C1 to C4 in the above-described examples, the acceptor is based on the structure of a known receptor. Other than calculating the information of the three-dimensional lattice points created in the inner pore of the body, extracting the functional group of the ligand molecule that interacts directly with the receptor, and setting the structural coordinates of the functional group It is possible to construct the molecular structure of the ligand by performing the same processing steps as in the examples.
[0068]
In addition, when the three-dimensional structure of the receptor alone has been elucidated and the environment of the drug binding site is known, instead of the steps R1 to R15 and C1 to C4 in the above-described examples, The same processing steps as in the embodiment, except that the information of the three-dimensional lattice points created in the above is calculated and the user sets the types and coordinates of two or more functional groups that can interact with the functional groups present in the receptor It is possible to construct the molecular structure of the ligand molecule by performing the above.
[0069]
[Example]
The method of the present invention was applied to an oncogenic promoter system consisting of TPA, teleocidin B-4, aplysiatoxin and thapsigargin. These oncogenic promoters are thought to bind to the same receptor, albeit with very different chemical structures. According to the method of the present invention, functional groups are extracted on the basis of the structures of the above four types of tumor promoters, and the extracted functional groups are connected by a connector, so that known compounds having an activity of these tumor promoters can be obtained. Once the molecular structure is constructed, the effectiveness of the method of the invention is demonstrated.
[0070]
The chemical structures of TPA, teleocidin, aprisiatoxin and thapsigargin are shown in FIG.
First, superposition of each molecule of TPA, teleocidin, aprisiatoxin, and thapsigargin was performed using RECEPS, and a receptor model was constructed based on the obtained superposition structure. As atomic coordinates of TPA, teleocidin, apriciatoxin, and thapsigargin, values obtained from the Cambridge Crystal Structure Database (CCSD) were used. The interval between the three-dimensional lattice points was set to 0.5 angstrom.
[0071]
In the step of R5 of RECEPS, automatic superposition is performed. In the step of R10, a superposition model was optimized by selecting a superposition structure having a good match between the molecular shape, electrostatic potential, and hydrophobic portion. Further, in the step of R13, an OH group (hydrogen donating functional group) and a C═O group (hydrogen accepting functional group) were selected as hydrogen bonding sites. FIG. 7 shows the three-dimensional structure of each molecule obtained by superposition using RECEPS. Moreover, the receptor model constructed | assembled by RECEPS is shown in FIG. The caged space shown in the diagram shows the space (receptor bore) that is assumed to be allowed by the ligand molecules. FIG. 8 also shows functional groups (C═O group and OH group) of the ligand molecule to which dummy atoms are added.
[0072]
Next, the molecular structure of the ligand was constructed by CONNECTOR. In the step of C4, an OH group (hydrogen donating functional group) and a C═O group (hydrogen accepting functional group) were extracted. In step C7, the limit value of the D value was set to 3 angstroms, and the limit value of the T value was set to 20 °. In step C8, tetravalent carbon, benzene ring, ethylene, amide (-NHCO-), and cyclohexane were prepared as fragment groups. In step C9, the rotation mode of the bond was set so that the rotatable single bond in the connector was rotated by 30 ° in the range of 0 to 360 °. In addition, in step C15, the atomic coordinates of ATOMhb1 and ATOMhb3 on the functional group side are the positions of the atoms corresponding to ATOMhb1 and ATOMhb3 in the TPA molecule among the molecules subjected to molecular superposition. .
[0073]
Note that the coordinates of the functional group were not changed in the step C21, and the functional group having the same interaction was not substituted in the step C33.
As a result, about 200 molecular structures were obtained. The molecular structure of one part is shown in FIG. These molecular structures have the same skeleton as TPA (Hecker, 1. Carcinogenesis: A Comprehensive Survey, Mechanisms of Tumor Promotion and Carcinogenesis, edited by TJ Slaga, A. Sivak and RR Boutwell, (1978) p.11-78). Tumor promoter activity confirmed by M. Ohno, Y. Endo, M. Hirano, A. Itai, K. Shudo (1993) Tetrahedron Letters, in press) Molecular structure B having a skeleton similar to ring benzolactam, Endo et al. (Oxygenated Cholesterol as Ligands for Cytosolic-Nuclear Tumor Promoter: Yakkasterol, Y. Endo, Y. Hashimoto, H. Fukasawa, K. Shudo, BBRC, 194, 1529-1935, (1993)) includes a molecular structure C having a skeleton similar to that of yakkasterol, which has been confirmed to bind to a receptor for an oncogenic promoter. Therefore, it is highly possible that other compounds and derivatives thereof exhibit an oncopromoter activity equivalent to or higher than that of TPA or the like.
[Brief description of the drawings]
FIG. 1 schematically illustrates the main steps in an example of the method of the present invention.
FIG. 2 shows a portion of a flowchart of one embodiment of the method of the present invention using CONNECTOR.
FIG. 3 shows a portion of a flowchart of an embodiment of the method of the present invention using CONNECTOR.
FIG. 4 shows a portion of a flowchart of an embodiment of the method of the present invention using CONNECTOR.
FIG. 5 shows a part of a flowchart of a molecular superposition method using RECEPS and receptor mapping.
FIG. 6 shows the chemical structure of teleocidin B-4, TPA, aprisia toxin and thapsigargin.
FIG. 7 shows the three-dimensional structures of teleocidin B-4, TPA, aprisiatoxin and thapsigargin obtained by superposition with RECEPS.
FIG. 8 shows a receptor model of an oncogenic promoter expressed using three-dimensional lattice points.
FIG. 9 shows a part of the molecular structure of a ligand for an oncopromoter receptor constructed by the method of the present invention.

Claims (7)

A method of constructing a molecular structure of the ligand having a physiological activity, were placed in (i) placing two or more functional groups in a position capable of interacting with receptors, the (ii) the step (i) A step of setting a combination of two atoms to be connected by a connector from atoms constituting a functional group, (iii) a step of creating a connector by combining two or more fragments, and creating a table storing these , (iv) from the table of the connector created in the process of the (iii), selecting a connector that connects between two atoms set the combination in step of said (ii), of (v) wherein (iv) rotating the rotatable bonds in the selected connector in step, step finding the proper conformation of the connector, and (vi) the between two atoms set the combination in step of said (ii) (iv seen in the process of) Which comprises the step, connecting the connector to take the proper conformation that. After the step (vi) , if a bond remains in the functional group, connect the functional group with the bond remaining to the atom to which hydrogen can be added with a new connector. The method according to claim 1, further comprising introducing a ring structure. The method according to claim 1, further comprising the step of introducing a ring structure by connecting a new connector between two atoms capable of adding hydrogen in the connector after the step (vi) . The method according to claim 1, wherein two or more functional groups are arranged at positions capable of interacting with a receptor by inputting atomic coordinates of a known ligand molecule and extracting the functional group from the molecular structure of the ligand. Input the atomic coordinates of two or more known ligand molecules, extract features common to the ligand molecules based on the superposition of the two or more ligand molecules, and extract functional groups from the molecular structure of the ligands based on it The method according to claim 1, wherein two or more functional groups are arranged at a position capable of interacting with a receptor . The two functional groups arranged in the step (i) are substituted with functional groups having equivalent physicochemical properties, and the steps (ii) to (vi) are performed. Method. the coordinates of the two functional groups which are arranged in the step (i) is replaced with the coordinates which is estimated to have equivalent effect and performing steps (ii) ~ (vi), in claim 1 The method described.

Images